human word 0.00335898
bleu bleu 0.00334256
word reordering 0.002990332
word alignment 0.002880422
source word 0.002698269
automatic word 0.002682731
similar word 0.002656444
first word 0.002626686
word order 0.002568012
target word 0.002566227
translation model 0.002557256
word level 0.002554462
bleu score 0.0024598989999999998
word alignments 0.002459055
measure word 0.002436152
much word 0.002434028
new word 0.0024207630000000003
word root 0.002417928
translation quality 0.002415859
large bleu 0.002398163
word tokens 0.00238065
word position 0.002365952
word sequence 0.002363498
hmm word 0.002351397
bleu tuning 0.002306329
bleu scores 0.002237655
translation task 0.00219439
bleu meteor 0.002164652
translation length 0.002153854
machine translation 0.002138712
small bleu 0.002129105
different system 0.002108568
wmt bleu 0.002066258
human evaluation 0.0020655960000000003
translation models 0.002060406
bleu brevity 0.00205207
bleu baseline 0.002049348
same system 0.0020460870000000003
bleu port 0.00201861
port bleu 0.00201861
translation hypothesis 0.00199579
hans bleu 0.001977584
bleu outputs 0.001967012
ibm bleu 0.001965993
ctb bleu 0.001955788
system tuning 0.001951319
bleu mtr 0.001948715
tune bleu 0.0019472270000000002
bleu qmean 0.001941204
system training 0.001908347
test data 0.001901757
evaluation metric 0.0018933959999999999
system level 0.001784122
system parameter 0.001780106
tuning metric 0.001735219
test set 0.001690357
translation 0.00164428
evaluation metrics 0.0016440489999999999
system types 0.0016330300000000002
large data 0.0016314580000000001
system outputs 0.001612002
human judgment 0.001608625
data set 0.00159775
human judgments 0.001594795
human evaluations 0.001585376
different translations 0.00157412
score translations 0.0015704410000000001
human ones 0.0015525530000000001
human preference 0.0015508820000000002
different systems 0.001545484
human judges 0.001537745
other metrics 0.001531288
other language 0.001528525
training data 0.001496652
distance reordering 0.001495447
evaluation set 0.001486401
metric lrscore 0.001459966
metric port 0.0014475
automatic metrics 0.001446944
tunable metric 0.001440906
possible score 0.001435619
other systems 0.001433651
automatic evaluation 0.001389347
metric yields 0.001386233
source words 0.001385248
news test 0.001382928
small data 0.0013624
combined metric 0.001359776
evaluation scores 0.001359601
chosen metric 0.0013592589999999998
intermediate metric 0.001358602
reference translations 0.001356679
source sentences 0.001347662
alignment task 0.001343922
evaluation task 0.001343336
test sets 0.0013433030000000001
same source 0.0013414759999999999
reordering models 0.001319848
system 0.00131627
other tuning 0.001315514
